INDEX

Explanations

personal pronouns or possessive pronouns

oai_token-act-pair · gpt-3.5-turbo

references to personal experiences or opinions expressed as questions and statements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ILCS

-0.87

ateral

-0.69

ï¸ı

-0.68

 Pole

-0.67

planes

-0.65

achus

-0.65

ioxide

-0.62

apixel

-0.60

axy

-0.60

 Waters

-0.60

POSITIVE LOGITS

 chose

0.74

 differed

0.73

 differs

0.68

 differ

0.67

 invoked

0.64

 sacrific

0.64

 practition

0.64

 staggered

0.63

 suspic

0.62

called

0.61

Activations Density 0.118%

personal pronouns or possessive pronouns

references to personal experiences or opinions expressed as questions and statements

No Comments

No Known Activations